Physics and Chemistry of the Earth 28 (2003) 943–956

www.elsevier.com/locate/pce

What is the limit of up-scaling rainwater harvesting in a river basin?

Stephen N. Ngigi
Department of Agricultural Engineering, University of Nairobi, P.O. Box 29053, Nairobi, Kenya

Abstract
The semi-arid savannah environment (SASE) of sub-Saharan Africa are characterized by low erratic rainfall which result to high
risk of droughts, intra-seasonal dry spells and frequent food insecurity. The main occupation is subsistence small-scale rainfed
agriculture and livestock production, which normally compete for the limited water resources. The main challenges to improving the
livelihoods of the small-scale farmers are how to upgrade rainfed agriculture to improve rural livelihoods and conserve nature, and
upgrade upstream landuse in balance with water needs for human and ecosystems downstream. There is an increased interest in
opportunities of improving rainfed agriculture through adoption of rainwater harvesting (RWH) technologies. However, there is
inadequate knowledge on hydrological impacts and limits of up-scaling rainwater harvesting at a river basin scale. Rainwater
harvesting has a potential of addressing spatial and temporal water scarcity for domestic, crop production, livestock development,
environmental management and overall water resources management is SASE. However, this potential has not been exploited
despite the occurrence of persistent low agricultural production and food shortage in sub-Saharan Africa. The need to quantify this
perceived potential and related hydrological impacts on a river basin led to the on-going research project titled ‘‘hydrological
impacts of up-scaling RWH on upper Ewaso Ng’iro river basin water resources management’’. It is envisaged that the study will
contribute to formulation of sustainable RWH up-scaling strategies to enhance food production and hydro-ecological balance in
semi-arid savannahs of Africa. This paper presents the preliminary findings of the study mainly focusing on assessment of the
potential of RWH technologies for improving food and water availability especially in semi-arid regions of eastern Africa. This was
achieved by evaluating six RWH case studies selected from four countries (Ethiopia, Kenya, Tanzania and Uganda). Despite the
success of a number of RWH systems, the rate of adoption is still low, hence making their impacts marginal. Nevertheless, there is a
knowledge gap on the limits of up-scaling RWH in a river basin, which the other components of the study will address. The as-
sessment of the hydrological impact of up-scaling RWH technologies is expected to provide answers to the question, what is the limit
of up-scaling rainwater harvesting in a river basin?

2003 Elsevier Ltd. All rights reserved.

Keywords: Rainwater harvesting; River basin; Fanya juu; Fanya chini; Food production; Water scarcity

1. Introduction interrelationships over longer time periods. These crises

1.1. Background
Most of the countries in sub-Saharan Africa (SSA)
are experiencing profound socio-economic and political
problems, the most dramatic being food crises and dis-
ruptive conflicts. The communities involved are experi-
encing a combination of both short-term, often acute
food crises, and long-term or chronic food shortages.
The former often translate into famine and starvation,
requiring emergency food aid. The latter are less obvi-
ous, for they are characterized by negative changes in
the economic, social and ecological factors and their
E-mail addresses: ngigi@gharainwater.org, gharp@warianchi.com
(S.N. Ngigi).
1474-7065/$ - see front matter
2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.pce.2003.08.015
threaten the stability and existence of the affected com-
munities and economies because their systems are ob-
viously failing to cope, increasing the vulnerability of the
people. A number of explanations have been advanced
for the endemic food insecurity in the SSA. Among
these, recurring drought and unreliable rainfall are the
most obvious. These include: adverse weather and
drought; rapid population growth rates that exceed rates
of food production; adoption of production systems
that accelerate environmental degradation and decline
in soil fertility; and retrogressive social organizations,
inadequate policies, legislation and institutional weak-
nesses.
Over 60% of the land in the SSA falls under semi-arid
savannah environments (SASE), where a majority of the
inhabitants are pastoralists although agro-pastoral and
farming communities have been slowing settling in these
944 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
areas due to population pressure in the high agricultural
potential areas. SASE is predominantly characterized by
low and variable rainfall, which rarely exceeds 800 mm,
with most areas receiving 200–350 mm annually. The
water resources are limited and poorly distributed.
There are few permanent rivers, and seasonal streams
that flow only during the wet season and remain dry for
the rest of the year. However, like the wetter regions,
SASE too is starting to experience land pressure re-
sulting due to population increase within the commu-
nities and also their livestock. This has significantly
raised pressure on pastures leading to overgrazing and
decline in vegetation cover in most of the areas. The
impact of the frequent droughts that hit the pastoral
areas has therefore been increasing over time. Huge
livestock numbers have been dying every time there is
drought (Kihara, 2002). Much of the pastureland has
lost grass cover and is often bare. This leaves the people
highly vulnerable. Consequently the ASALS form spe-
cific pockets of poverty and food insecurity, and ensuing
conflicts, especially over diminishing natural resources––
mainly water and pasture.
Nevertheless, unreliable rainfall and low soil fertility
has continued to threaten food production in the SSA
thus making food security a major concern. Currently,
vast areas of SSA are facing drought and the threat of
famine despite the fact that overall food production
could be adequate. Relief food has on many occasions
saved lives in the region from severe famine situations.
Food relief will continue to be required as long as
transportation facilities are poor and local food pro-
duction in drought prone areas is inadequate. Given the
poor transportation infrastructure, emphasis on local
food production appears the most logical approach to
improved food security.
Agriculture is the major economic activity for the
countries of the SSA, engaging between 75% and 85% of
the people of those countries. Consequently, it is
strongly underscored that agriculture is the backbone of
these countries’ economic development and their peo-
ple’s well being in the foreseeable future. A survey of 277
societies in sub-Saharan Africa by Hunt-Davis (1986)
showed that approximately 86% depended primarily on
agriculture, 6% on animal husbandry, and animal hus-
bandry and agriculture are co-dominant for another 3%.
Of the rest, 2% rely primarily on fishing; 1% on fishing
and agriculture equally, and some minorities on hunting
and gathering. Thus the livelihood in this region is based
on small-holder rural agriculture, with low levels of
productivity and simple tools, making them over-
dependent on the status of the natural environment. Sea-
sonal rainfall dominates the lives of most of the people,
as it determines their activities geared towards earning a
livelihood based on exploitation of the resources of the
land. Duckham and Masefield (1985) stated that in the
tropics generally, rainfall is the main determinant of
agricultural activities. The same fact had been expressed
by Jodha and Mascarenhas (1985) as characteristic of
much of the rest of Africa. In Greater Horn of Africa
(GHA), rainfall––amount, timing, duration and distri-
bution––was identified by subsistence farmers as the
dominant determining factor for food production and
security.
Therefore, the problems related to food security and
recurrent famine need urgent solutions, especially in the
SASE, where environmental degradation has further
decreased agricultural productivity, making inhabitants
even more susceptible to drought and other natural di-
sasters. Unless sustainable food production technologies
are adopted, alleviation of poverty and food security
will remain elusive. One promising technology for rural
landuse systems is rainwater harvesting. This is the
process of interception and concentration of runoff and
its subsequent storage in the soil profile or in artificial
reservoirs for crop production. The process is distin-
guished from irrigation by three key features: the
catchment area is contiguous with the cropped area and
is relatively small; the application to the cropped area or
reservoir is essentially uncontrolled; and water harvest-
ing can be used for purposes other than crop produc-
tion. There are many techniques being used to enhance
crop production in the ASALs of the SSA. However, the
viability of these solutions needs to be evaluated in re-
lation to environmentally sustainable factors, climatic
conditions, soil characteristics, farming systems and
socio-cultural and gender perspectives in which they are
practiced.
Participatory evaluation is needed to determine viable
options and adaptive strategies for sustainable food
production in the ASALs. Needless to say, the solutions
must be landuser-oriented, hence the need for a partic-
ipatory technology development approach. The project
gave special attention to RWH technologies and sys-
tems, which are being used by land users. Any activity to
improve on landusers’ innovations and the applicability
of those innovations will be a major contribution to
food security in this famine prone region. It is with a
sense of urgency that one notes the relevance of RWH
technologies in certain limited but significant areas of
Africa, both for food production and for soil and water
conservation (Pacey and Cullis, 1986). Despite this re-
alization, little practical information exists on RWH
technologies, which can be applied on site specific situ-
ations. RWH is one of the approaches to integrated land
and water management, which could contribute to re-
covery of agricultural production in dry areas.
1.2. Food production and water scarcity in sub-Saharan
Africa
The semi-arid areas of SSA are characterized by low
annual rainfall concentrated to one or two short rainy
 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
seasons. The average annual rainfall varies from 400 to
600 mm in the semi-arid zone, and ranges between 200
and 1000 mm from the dry semi-arid to the dry sub-
humid zone (Rockstr€ om, 2000). The length of the
growing period ranges from 75 to 120 and 121–179 days
in the semi-arid zone and dry sub-humid zone respec-
tively. Potential evaporation levels are high, ranging
from 5 to 8 mm/day (FAO, 1986) giving a cumulative
evapotranspiration of 600–900 mm over the growing
period. This explains the persistence water scarcity
coupled with low crop yields. Water scarcity could also
be attributed to poor rainfall partitioning leading to
large proportion of non-productive water flows––not
available for crop production. The nature and occur-
rence of rainfall in SASE of SSA provides more insight
in the food production and water scarcity situations.
Rainfall is highly erratic, and normally falls as in-
tensive storms, with very high intensity and spatial and
temporal variability. The result is a very high risk for
annual droughts and intra-seasonal dry spells (Rock-
str€
om, 2000). From past experience, severe crop reduc-
tions caused by dry spells occurs 1–2 out of 5 years,
while total crop failure caused by annual droughts that
occur once in every 10 years in semi-arid SSA. This
means that the poor distribution of rainfall, more often
than not, leads to crop failure than absolute water
scarcity due to low cumulative annual rainfall. Unfor-
tunately, most dry spells occur during critical crop
growth stages (this explains frequent crop failure and/or
low yields), and hence the need of dry spell mitigation by
improving water productivity in SSA.
From the above brief overview of rainfall patterns,
there is a growing understanding that the major crop-
ping systems in SSA are not sustainable (Benites et al.,
1998), hence the persistence low food production (food
shortage) and reliance on food relief. Nevertheless,
majority of the population in SSA make their meager
living from rainfed agriculture, and depend on to a large
extent on smallholder, subsistence agriculture for their
livelihood security (e.g., Botswana, 76%; Kenya, 85%;
Malawi, 90%; and Zimbabwe, 70–80% of the population
(Rockstr€ om, 2001). Moreover, an estimated 38% of the
population in SSA roughly 260 million people lives in
drought prone SASE (UNDP/UNSO, 1997). This may
explain why most of the population is poor––rely on
unsustainable farming systems, majority living below
the poverty limit (<US$ 1 per day). The key role of
agriculture in Africa’s economic life is apparent––agri-
culture accounts for 35% of the continent’s GDP, 40%
of its export, 70% of its employment, and more than
70% of the population depend for their livelihoods on
agriculture and agri-business (Kijne, 2000).
The problem of low food production is further ag-
gravated by limited new land for cultivation, land and
environmental degradation, poor infrastructure, politi-
cal and social crises, bad governance, insecure land
945
tenure, health/diseases outbreak, inadequate knowledge/
capacity, and donor dependency syndrome. Thus the
ever increasing food demand and household income
needed in SSA have to be achieved through an increase
in biomass produced per unit land and unit water
(Rockstr€om, 2001). In the past, very little attention has
been paid to the development of rainfed agriculture in
SSA except provision of conventional irrigation pro-
jects. However, most of these projects have proven (e.g.
Bura irrigation scheme in Kenya) to be unnecessary,
costly and environmentally unsustainable. Hence the
need to focus on opportunities of increasing efficiency of
limited water in rainfed, smallholder agriculture in the
SASE of SSA. Otherwise feeding the ever growing
population (at a rate of 2–3% per year) with diminishing
crop yields (oscillating around 1 ton/ha for food grains)
in SSA (Rockstr€ om, 1999) will remain elusive, and
current generation’s biggest challenge.
1.3. Rainwater harvesting technologies and systems
Rainwater harvesting which is broadly defined as the
collection and concentration of runoff for productive
purposes (crop, fodder, pasture or trees production,
livestock and domestic water supply, etc.), has ancient
roots and still forms an integral part of many farming
systems worldwide (Evanari et al., 1971; Shanan and
Tadmor, 1976; Critchley, 1987; Critchley and Siegert,
1991; Agarwal and Narain, 1997). It includes all meth-
ods of concentrating, diverting, collecting, storing, and
utilizing and managing runoff for productive use.
However, in situ systems i.e. on-farm/cropland water
conservation––to enhance soil infiltration and water
holding capacity––dominate, while storage systems for
supplemental irrigation are less common, especially in
SSA (SIWI, 2001). Nevertheless, a recently concluded
evaluation of RWH in four GHA countries (i.e. Ethio-
pia, Kenya, Tanzania and Uganda) revealed that, de-
spite the relatively high investment costs compared to
in situ systems, RWH for supplemental irrigation is slowly
being adopted with high degree of success (Kihara,
2002). In this system, surface runoff from small catch-
ments (1–2 ha) or adjacent road runoff is collected and
stored in manually and/or mechanically dug farm ponds
(50–1000 m3 storage capacity). Due to the low volumes
of water stored compared with crop water requirements,
improved benefits of these systems are derived by in-
corporating efficient water application methods such as
low pressure (0.5–1.5 m) drip irrigation (Ngigi et al.,
2000; Ngigi, 2001).
Furthermore, on-farm research in semi-arid locations
in Kenya (Machakos district) and Burkina Faso (Ou-
agouya) during 1998–2000 (Barron et al., 1999; Fox and
Rockstr€ om, 2000) indicates a significant scope to im-
prove water productivity in rainfed agriculture through
supplemental irrigation, especially if combined with soil
946 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
fertility management. The results were more promising
on soils with higher water holding capacity on which
crops seem to cope better with intra-seasonal dry spells.
However, incremental water productivity improvements
are only achieved during rainy seasons with severe dry
spells, while rainy seasons with adequately distributed
rainfall the incremental value can be negative (Rock-
str€om et al., 2001).
Runoff is collected mainly from ground catchments as
well as ephemeral streams (flood water harvesting) and
road/footpath drainage. The storage is either in different
structures (tanks, reservoirs, dams, water pans, etc.),
mainly for supplemental irrigation systems, or soil
profile (for in situ and flood irrigation). RWH can be
considered as a rudimentary form of irrigation (Fentaw
et al., 2002). The difference is that with RWH the farmer
has no control over timing, as runoff can only be har-
vested when it rains. In regions where crops are entirely
rainfed, a reduction of 50% in the seasonal rainfall, for
example, may result in a total crop failure (Critchley and
Siegert, 1991). However, if the available rain can be
concentrated on a smaller area, reasonable yields will
still be received. Fig. 1 shows the principle of RWH,
which is common for different classifications, except
in situ (no runoff) systems which capture rainfall where
it falls. Classification of runoff-based RWH technologies
depends:
• Source of runoff (external) or within-field catchments;
• Methods of managing the water (maximizing infiltra-
tion in the soil, storing water in reservoirs and inun-
dating cropland with floods); and
• Use of water (domestic, livestock, crop production,
gully rehabilitation, etc.).
RWH systems operate at different scales (household,
field and catchment/basin), and can affect water avail-
Catchment
(natural surfaces, roads/footpaths, gullies, rills
ephemeral streams, croplands, pasture,
hillslopes)
Runoff
Conveyance
and/or
“Storage
Cropland
(water applied directly or through
irrigation for storage systems)
Fig. 1. The principle of runoff-based rainwater harvesting technology.
ability and management for downstream and natural
ecosystems like wetlands and swamps, due to reduced
catchment water yields. Therefore, even though RWH
practices can be efficient in increasing the soil moisture
for crops (principle objective) in water scarce areas, each
technique has a limited scope due to hydrological and
socio-economic limitations. Rockstr€ om (2000) high-
lighted the major hydro-climatic hazards in SASE
farming as
• Poor rainfall partitioning, where only a small fraction
of rainfall reaches the root zone, coupled with within-
field crop competition for soil water;
• The high risk of periods of below optimal cumulative
soil water availability during the growth season (i.e.
not necessarily dry spells, but rather situations when
soil water availability is below crop water require-
ments for optimal yields due to low cumulative rain-
fall levels); and
• The high risk of intermittent droughts, or dry spells,
occurring during critical crop growth stages (i.e. not
necessarily a lack of cumulative soil availability, but
rather periodic water stress due to poor rainfall distri-
bution.
2. Rainwater harvesting technologies and systems in SSA
This section briefly presents the different RWH
technologies and systems found in SSA, focusing on
their classifications, and their opportunities and limita-
tions for improving rainfed agriculture in SASE. The
term RWH is used in different ways and thus no uni-
versal classification has been adopted. However, ac-
cording to Oweis et al. (1999) the following are among
its characteristics: RWH is practiced in ASALs where
surface runoff is intermittent; and is based on the utili-
zation of runoff and requires a runoff producing area
(catchment) and a runoff receiving area (cropped area
and/or storage structures). Therefore, each RWH sys-
tem, except in situ water conservation (as shown in Fig.
1 above) should have the following components: runoff
producing catchment, runoff collection (diversion and
control) structures, and runoff storage facility (either soil
profile in cropland or distinct structure (farm ponds,
tanks, water pans, earthdams, sand dams, sub-surface
dams, etc.).
To avoid further confusion, and facilitate the pre-
sentation of various types of RWH technologies and
systems, the classification shown in Fig. 2 below is based
on runoff generation process, type of storage/use and
size of catchments is adopted. The runoff generation
criteria yields two categories––runoff farming (where
runoff is generated i.e. runoff-based systems) and in situ
water conservation (rainfall conserved where it falls).
The runoff storage criteria also yields two categories––
 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956 947

RWH SYSTEMS

In-situ Water Conservation

Runoff-based Systems
(Tillage and cultural practices)
(Catchment and/or storage)

Direct Application Systems Storage Systems

(Runoff diversion into (Distinct storage structures
cropland where soil profile for supplemental irrigation
provide moisture storage) and other uses)

Micro-Catchment Systems Small Catchment Systems Macro-Catchment Systems

(Within field/internal (Runoff generated from small (Flood diversion and
catchments systems) external catchments and spreading i.e. spate
diverted to cropland/pasture) irrigation)

Fig. 2. Adopted classification of rainwater harvesting technologies and systems in sub-Saharan Africa.

soil profile storage (direct runoff application) and dis-
tinct storage structures for supplemental irrigation,
livestock, domestic or commercial use). Whilst the size
of catchment criteria yields three categories––macro-
catchments (flood diversion and spreading i.e. spate
irrigation), small external catchments (road runoff, ad-
jacent fields, etc.), and micro (within field)-catchments
(e.g. Negarims, pitting, small bunds, tied ridges, etc.).
Moreover, runoff storage structures capture runoff
mainly from small catchments especially for smallscale
landusers, but macro-catchments with large storage
structures could also be used for large-scale or com-
munity-based projects. In situ water conservation could
also be considered under soil profile storage systems,
only that in that case direct rainfall is stored, but not
surface runoff. However, the classification is further
complicated by the fact that a number of RWH tech-
nologies are integrated or combined by landusers, for
example, fields under conservation tillage in Laikipia
district also incorporate runoff spreading from small
external catchments such as road/footpath drainage and
adjacent fields. It is also common to find runoff from
external catchment being directed into cropland with
farm ponds for supplemental irrigation. In situ water
conservation is also combined with runoff farming on
farms with terraces, in which the terrace channel (mainly
fanya juu and contour ridges/bunds) collects and stores
runoff from small external catchments while the crop-
land between the channels harvest and conserve direct
rainfall. However, excess runoff that may be generated
from the cropland between the terrace channels would
be collected at the channel.
The following sub-sections highlight some of the
RWH technologies and systems that have been tried,
experimented and practiced in different parts of sub-
Saharan Africa, in addition to those identified and
evaluated as part of the GHARP case studies in parts of
Ethiopia, Kenya, Uganda and Tanzania.
948 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
2.1. In situ rainwater conservation
In situ rainwater conservation technologies are dis-
tinct from runoff farming systems in that they do not
include a runoff generation area, but instead aims at
conserving the rainfall where it falls in the cropped area
or pasture. The most common technology is conserva-
tion tillage which aims to maximize the amount of soil
moisture within the root zone. A number of cultural
moisture practices such as mulching, ridging, addition of
manure, etc. could fall under this category. Small field/
farm structures such as tied ridges/bunds within cropped
area that conserve direct rainfall without Ôexternal’––
outside cropland boundary, i.e. no distinct catchment
area, except overflow from upstream sections also falls
under this category. Within cropland or pasture contour
bunds/ridges, bench terraces, and sweet potatoes ridges
practiced in Rakai district of Uganda could also fall
under this category.
In situ rainwater conservation technology is one of
the simplest and cheapest and can be practiced in almost
all the landuse systems. In situ water conservation sys-
tems are by far the most common (Rockstr€ om, 2000)
and are based on indigenous/traditional systems (Reij
et al., 1996; LEISA, 1998). The primary objective has
been to control soil erosion and hence manage the
negative side-effects of runoff––soil and water conser-
vation measures, i.e. ensures minimal runoff is gener-
ated. The positive effect of in situ water conservation
techniques is to concentrate within-field rainfall to the
cropped area. In a semi-arid context, especially with
coarse-textured soil (especially sandy soils common in
the ASALs) with high hydraulic conductivity, this
means that in situ conservation may offer little or no
protection against the poor rainfall distribution. In such
cases, the farmers will continue to live at the mercy of
the rain. In effect, the risk of crop failure is only slightly
lower than that without any measures. However, soil
improvements such as addition of manure would en-
hance realization of better yields.
2.1.1. Conservation tillage
Conservation tillage is defined as any tillage sequence
having the objective to minimize the loss of soil and
water, and having an operational threshold of leaving at
least 30% mulch or crop residue cover on the surface
throughout the year (Rockstr€ om, 2000). However, with
respect to small-scale farmers in SASE, conservation
tillage is defined as any tillage system that conserves
water and soil while saving labour and traction needs.
Conservation tillage aims at reversing a persistent trend
in farming systems of reduced infiltration due to com-
paction and crust formation and reduced water holding
capacity due to oxidation of organic materials (due to
excessive turning of the soil). From this perspective,
conservation tillage qualify as a form of water harvest-
ing, where runoff is impended and soil water is stored in
the crop root zone (Rockstr€ om et al., 1999).
Unlike the conventional tillage systems, based on soil
inversion which impedes soil infiltration and root pene-
tration, conservation tillage covers a spectrum of non-
inversion practices from zero-tillage to reduced tillage
which aim to maximize soil infiltration and productivity,
by minimizing water losses (evaporation and surface
runoff) while conserving energy and labour. Kihara
(2002) revealed the successes of conservation tillage in
harnessing rainwater and improving yields. Field visits
in Machakos revealed that, during the recent below
average short rains (2001/2002), farms where conserva-
tion tillage was practiced had good harvest while
neighbouring farms without convention tillage had lit-
erally no harvest––conspicuous contrast.
Conservation tillage has several attractive effects on
water productivity (Rockstr€ om et al., 2001) compared to
traditional soil and water conservation systems such as
fanya juu terracing in Machakos district. In addition to
enhancing infiltration and moisture conservation, it en-
ables improved timing of tillage operations, which is
crucial in semi-arid rainfed farming. It can also be ap-
plicable on most farmlands compared say to storage
RWH systems for supplemental irrigation. Promotion
of animal drawn conservation tillage tools such as rip-
pers, ridgers and sub-soilers among smallholder farmers
in semi-arid Machakos and Laikipia districts (Kenya)
has resulted in significant water productivity and crop
yields (Kihara, 2002; Muni, 2002). There are many
documented examples of successful conservation tillage
practices in ESA, where crop yields have been increased
through the conservation of soil water and nutrients
and/or draught power needs have been reduced (Rock-
str€om et al., 1999).
The findings of the case studies in Laikipia and
Machakos districts of Kenya reveals that conservation
tillage (sub-soiling and ridging) have improved yields by
more than 50% (Kihara, 2002; Muni, 2002). The po-
tential of conservation tillage is tremendous especially
with communities already using animal drawn imple-
ments for their tillage operations. This is because con-
servation tillage implements are compatible with the
conventional tools. On large scale farming systems in
Laikipia district, tractor drawn conservation tillage im-
plements have improved wheat yields. Pastoral com-
munities are also not being left behind, as ground
scratching using animal or tractor drawn tools have
improved pasture development in Laikipia district. In
Dodoma, Tanzania, trench cultivation, a form of con-
servation tillage have been developed by innovative
farmers, where shallow trenches are dug, filled with or-
ganic materials then covered by soil to form ridges on
which crops are planted (Lameck, 2002). The Ôorganic’
furrows between the ridges capture water, which seeps
into the covered trenches and is slowly extracted by the
 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
crops. The organic material improves soil fertility and
water holding capacity. This seems to be an improve-
ment of the furrow and ridge systems as used in Kitui
and Machakos. The furrows and ridges are made using
animal drawn mould board ploughs. Seeds are planted
in the furrow, which collects water between the ridges.
After seedlings develop, weeding operation (using ani-
mal drawn ridgers) ensures that the furrows and ridges
alternate––the crops grow on the ridges while the fur-
rows captures and concentrates the rainwater. In trench
cultivation, the ridges and furrows are rotated after
each growing season and have enhanced crop yields in
otherwise low yielding areas.
2.2. Runoff farming
The runoff farming systems, which entail runoff gen-
eration either within field or from external catchments
and subsequent application either directly into the soil
profile or through periodic storage for supplemental ir-
rigation, are classified according to two criteria (as
shown in Fig. 2 above): runoff storage and/or applica-
tion, and size of catchments.
2.2.1. Storage RWH systems
RWH systems with storage for supplemental irriga-
tion are becoming popular in semi-arid districts of
Kenya (e.g., Machakos, Laikipia, and Kitui). They have
also been introduced in Ethiopia (near Nazareth) on
experimental basis by RELMA. Moreover, small stor-
age systems are all over parts of Ethiopia (e.g. Tigray),
and other places around Africa (Fentaw et al., 2002).
Initial results from RWH experiments in Machakos
district, which focused on the feasibility of using earth-
dams for supplemental irrigation of maize have been
encouraging (Rockstr€ om et al., 2001). The main chal-
lenge with this initiative is to assess whether it is possible
to design simple and cheap earthdams or farm ponds
that could permit gravity-fed irrigation to reduce the
cost of lifting water.
In the semi-arid parts of Laikipia district (Kenya),
underground water tanks (50–100 m3 capacity) have
been promoted mainly for kitchen gardening. The tank
surfaces are usually sealed with polythene lining, mor-
tar, rubble stones or clay to reduce seepage losses while
covering the tanks, with either local material (thatch or
iron sheet), minimizes evaporation. However, similar
initiative in Kitui district was discouraging as most of
the mortar sealed underground tanks ended up cracking
and hence being abandoned (Ngure, 2002). In Laikipia,
loss of water through seepage has been identified as a
major drawback (Kihara, 2002). Thus despite the posi-
tive impact realized by this technology, its widespread
adoption could be hampered if simple seepage control
measures are not devised. Concrete sealing seem to work
well in Ng’arua division of Laikipia district, but the cost
949
may be beyond the reach of many farmers. Farmers are
still experimenting with various seepage control meth-
ods, among them, plastic lining (found not durable),
butimen lining, clay lining and even goats trampling.
Nevertheless, seepage control still remains a major
challenge.
Other storage systems used by smallscale farmers in
semi-arid districts of eastern Kenya are rock catch-
ments/dams, sand dams and sub-surface dams (Gould
and Petersen, 1991; Pacey and Cullis, 1986). Sand dams
and sub-surface dams are barriers constructed along
sandy riverbeds––a common phenomenon in most semi-
arid environments in GHA––to retain water within the
trapped sand upstream. These systems have provided
water for decades especially in Machakos and also in
some parts of Kitui district. They have also been in-
troduced in the Dodoma area of Tanzania but their
potential has however, not been realized. They provide
water for all purposes and could lead to environmental
improvement, for example in Utooni in Machakos and
some parts of central Kitui. The impacts of sand dams
on food security have been highlighted by Isika et al.
(2002). They are mainly used for domestic purposes, but
in several cases also used for smallscale irrigation
(Rockstr€ om, 2000). Rock catchment dams are masonry
dams, for capturing runoff from rock surfaces/catch-
ments, with storage capacities ranging from 20–4000 m3 .
They are generally used for domestic purposes, but can
also be used for kitchen gardening, for example in Kitui
district (Ngure, 2002).
RWH storage systems offers the landuser a tool for
water stress control––dry spell mitigation. They reduce
risks of crop failures, but their level of investment is high
and requires some know-how especially on water man-
agement. However, these systems also to some extent
depend on rainfall distribution. During extreme drought
years, very little can be done to bridge a dry spell oc-
curring during the vegetative crop growth stage if no
runoff producing rainfall have fallen during early growth
stages. Under normal intra-seasonal droughts, the farmer
will be assured of a better harvest and hence it is worthy
any investment to improve crop production in the semi-
arid tropics of SSA. Nevertheless, location of the storage
structure with respect to cropland needs to be addressed.
Conventionally, the reservoirs are located downstream,
thus requiring extra energy to deliver the water to the
crops. However, it would be more prudent to locate the
reservoir upstream of the cropland to take advantage of
gravity to deliver the water (Rockstr€ om, 2000).
Runoff is collected from grazing land, uncultivated
land, cultivated land and road drainage and directed
into small manually constructed reservoirs (50–200 m3 ).
The stored water is mainly utilized for kitchen gardening
and establishment of orchards. This technology was
introduced in Laikipia district Kenya in the late 1980s
by the Anglican Church of Kenya and has shown
950 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
promising results. It has been promoted by Dutch-sup-
ported ASAL and SARDEP progammes with limited
success due to seepage related problems. Various reme-
dies are being tried to reduce seepage to realize maxi-
mum benefits from this technology. In Kenya, it has also
been introduced in Machakos district by RELMA.
Optimal benefits could be realized if appropriate water
lifting and application technologies such as treadle
pump and drip irrigation are incorporated. Farm ponds
have also been used for watering livestock. At commu-
nity level, earth dams or water pans are constructed
to store large quantities of water, especially for live-
stock and small-scale irrigation. These water pans and
earth dams are the lifeline for livestock in the ASAL of
Kenya, Somalia and parts of Uganda (southern and
northeastern). The earth dams were introduced by white
settlers while the water pans have been traditional
sources of water e.g. hafirs (water pans) in northeastern
province of Kenya, parts of Somalia and western Sudan
(Critchley, 1987). Concrete/mortar lined underground
tanks (100–300 m3 ) are used for domestic and some
livestock (milking cows, calves or weak animals, sepa-
rated from the main herds) in Somaliland (Pwani, 2002).
2.2.2. Direct runoff application systems
This category of RWH technology is characterized by
runoff generation, diversion and spreading within the
cropland, where the soil profile acts as the moisture
storage reservoir. This technology is further classified,
according to size of catchments: macro-catchments
systems––large external catchments producing massive
runoff (floods) which is diverted from gullies and
ephemeral streams and spread into cropland, i.e. spate
irrigation; small external catchments (e.g. road drainage,
adjacent fields, etc.) from which runoff is diverted into
cropland; and micro-catchments normally within crop-
land which generate small quantities of runoff for single
crops, group of crops or row crops.
2.2.3. Flood diversion and spreading (spate irrigation)
systems
Flood diversion and spreading (i.e. spate irrigation)
refers to RWH system where surface runoff from macro-
catchments concentrating on gullies and ephemeral
streams/water courses is diverted into cropped area and
distributed through a network of canals/ditches or wild
flooding and subsequently retained in the field by bunds/
ridges. It entails controlled diversion of flash floods from
denuded highlands to cropped land well prepared to
distribute and conserve the moisture within the plants
rootzone. The rainfall characteristics in the semi-arid
savannah environment occurs as high intensity storms
that generate massive runoff that disppear within a short
period through seasonal waterways. Worse still the
number of rainstorms are normally limited within the
short rainy seasons.
Extensive flood irrigation of paddy rice in cultivation
basins (commonly referred to as ‘‘majaluba’’) created
from 25–100 cm high earth bunds, is practiced in semi-
arid central parts of Tanzania (Dodoma, Singida and
Shinyanga) (Mwakalia and Hatibu, 1992; Hatibu et al.,
2000; Lameck, 2002). It is estimated that 32% of Tan-
zania’s rice production originate from cropland where
RWH is practiced. Similar techniques have been used
for maize and sorghum in Tanzania.
Spate irrigation in northern Ethiopia and Eritrea,
involve capturing of storm floods from the hilly terrain
and diversion into leveled basins in the arid lowlands
croplands. In Kobo Wereda (south of Tigray), spate
irrigation is well developed with main diversion canals,
secondary/branch, tertiary and farm ditches which dis-
tribute flood water into cultivation basins with contour
bunds to enhance uniform water application. A series of
main canals for different group of farmers are normally
constructed together to reduce destruction by floods
(Fentaw et al., 2002). Farmers in Kobo plains in
northern Ethiopia have developed a traditional irriga-
tion system that diverts part of such floods to their
farms. These system have sustained livelihoods that
would otherwise be impossible in that dry part of the
country. These systems are similar in principle to those
developed by the early settlers of the Negev Desert in
Israel. The system has also been tried in Konso, south-
ern Ethiopia. This technology has also been practiced in
Turkana district, Kenya for sorghum production and
parts of Sudan (Cullis and Pacey, 1992). In western
Sudan, terraces and dykes are used for spreading runoff
from wadis onto vertisols (Critchley, 1987). The poten-
tial of these systems are enormous and if improved and
promoted could lead to food security.
The use of external catchments for runoff collection
immediately adds water to the field scale water balance.
With flood irrigation systems in the SASE where abso-
lute crop water scarcity is common, crop yields can be
improved substantially during years with reasonably
good rainfall distribution. The farmers still live under
the mercy of the rains, but when it rains, the supply of
water to the root zone exceeds rainfall depths. This
disparity can be addressed by introducing storage fa-
cilities.
2.2.4. Small external catchment systems
These include a form of smallscale flood/runoff di-
version and spreading either directly into cropland or
pasture through a series of contour bunds or into terrace
channels and other forms of water retention structures.
The runoff is either conveyed through natural water-
ways, road drainage or diversion/cutoff drains. Road/
footpath runoff harvesting is practiced in parts of Kenya
(Machakos and Laikipia), in which flood water from
road/footpath drainage is diverted either into storage for
supplemental irrigation or into croplands (wild flooding,
 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
contour bunds, deep trenches with check-dams to im-
prove crop yields. Similar system is practiced in south-
western Uganda, where runoff from gullies, grazing
land, or road drainage is diverted into banana planta-
tions (Kiggundu, 2002).
Fanya juu terraces which were previously used with
diversion/cutoff drains for soil conservation especially in
Machakos and Kitui have been adopted for rainwater
harvesting. They are modified by constructing planting
pits mainly for bananas and tied ridges (check dams) for
controlling the runoff. The outlet is blocked to ensure as
much runoff as possible is retained while spillways are
provided to discharge excess runoff, which is normally
diverted into the lower terraces. Runoff spreading has
also been accomplished by contour bunds in Laikipia
district. They collect and store runoff from various
catchments including footpaths and road drainage. The
stored runoff seeps slowly into lower terraces ensuring
adequate moisture for crops grown between the terrace
channels. In southern Uganda, a similar system has been
adopted, in which contour ridges/bunds, (shallow fanya
juu terraces) tied at regular intervals are used in banana
plantations. The runoff from hilly grazing lands is dis-
tributed into the banana plantations by contour ridges.
Agroforestry (for firewood and fodder) is also incorpo-
rated, where trees planted on the lower side and Napier
or giant Tanzania grass along the ridges. This system has
tremendously improved the yield of the bananas and has
enhanced zero grazing. Contour ridges and infiltration
trenches have also been adopted to improve pasture in
southern Uganda (Kiggundu, 2002). The infiltration
trenches are dug at specified intervals according to the
land slope and tied at regularly intervals to allow water
retention and subsequent infiltration. The soil is either
thrown upward (fanya juu) or downwards (fanya chini)
and stabilizing grass or fodder crops. Runoff from uphill
catchments is normally diverted into these contour dit-
ches (infiltration trenches) to increase soil moisture.
In eastern part of Sudan, a traditional system of
harvesting rainwater in ‘‘terraces’’ is widely practiced
(Critchley, 1987). It consists of earthen bunds with wing
walls which impound water to depths of at least 50 cm
on which sorghum is planted. Within the main bund
there may be smaller similar bunds which impound less
runoff on which planting can be done earlier.
2.2.5. Micro-catchment systems
This involve runoff generation within the farmer’s
field and subsequent concentration on either a single
crop especially fruit trees, a group of crops or row crops
with alternating catchment and cropped area mainly
along the contours. A number of within-field RWH
systems fall under this technology, in which crop land is
subdivided into micro-catchments that supply runoff
either to single plants (e.g. pawpaw or oranges) for ex-
ample Negarims in Kitui, Kenya or a number of plants
951
(e.g. maize, sorghum etc) in case of chololo pits in
Dodoma, Tanzania. Pitting techniques, where shallow
planting holes (<25 cm deep) are dug for concentration
of surface runoff and crop residue/manure, are found in
many farming systems throughout SSA. They come in
many names and include zai pits (Burkina Faso), mat-
engo pits (southern highlands of Tanzania) and tum-
bukiza for napier grass and banana or pawpaw pits
(Kenya). Moisture retention terraces and ditches are
other micro-catchment techniques promoted and adop-
ted in SESA. The following are more examples:
• Fanya juu terraces, which are made by digging a
trench along the contour, and throwing the soil up-
slope to form an embankment. They have made a very
significant impact in reducing soil erosion in semi-
arid areas with relatively steep slopes (Thomas,
1997; Tiffen et al., 1994). They have been used for
RWH by incorporating tied ridges in the channel
with closed outlets.
• Fanya chini, in which the soil is thrown downslope in-
stead of upslope, was developed in Arusha region,
Tanzania.
• Contour bunds, e.g. stone bunds and trashlines in dry
areas of southern Kenya and retention ditches and
stone terraces in Ethiopia. Yields of sorghum are re-
portedly increased by up to 80% using contour bunds
in northwestern Somalia (Critchley, 1987).
• Micro-basins, which are roughly 1.0 m long and <50
cm deep, are often constructed along the retention
ditches for tree planting (e.g., northern Tigray, Ethi-
opia) (Lundgren, 1993). Sweet potato ridges/bunds in
southern Uganda fall under this category (Kiggundu,
2002). In Kwale districts of Kenya, tied ridges and
small basins have been reported to improve maize
yields by more than 70%.
• Semi-circular earth bunds (demi-lunes) are found in
ASALs for both rangeland rehabilitation and for an-
nual crops on gently sloping lands (e.g. Baringo and
Kitui districts) (Thomas, 1997). Semi-circular bunds
adopted for establishment of tree seedlings in de-
nuded hilly areas in southern Uganda applies the
same principle (Kiggundu, 2002).
• Negarims micro-catchment are regular square earth-
dams turned 45 from the contour to concentrate sur-
face runoff at the lowest corner of the square (Hai,
1998) are found in eastern province of Kenya.
• Large trapezoidal bunds (120 m between upstream
wings and 40m at the base) have been tried in arid
areas of Turkana district, northern Kenya for sor-
ghum, trees and grass (Thomas, 1997).
• Infiltration trenches/ditches, which are dug along the
contour, at specified intervals according to the slope,
for retaining runoff in banana plantations in south-
western Uganda, Mbarara and Rakai districts (Kigg-
undu, 2002).
952 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
• Circular depressions (3–4 m in diameter and <1.0 m
deep) where a variety of crops are inter-cropped
and literally allows no runoff from the fields are prac-
ticed in southern Ethiopia.
3. Role of rainwater harvesting technologies
3.1. Potential for improving food production
There are a number of promising interventions for
improving water availability either for crop production
or other uses in the dry parts of the SSA region. A few
techniques especially for irrigation have been tested and
proven successful but majority, which are mainly land-
users’ innovations remain unproven. It is evident that the
introduction of new technologies without landusers
participation, however novel they may be, has not been
successful. One such project is the multi-million Bura
irrigation scheme in Kenya. On the other hand, the
landusers’ ingenuity has certainly paid dividends. The
challenge now is to evaluate landusers innovations and
traditional systems to determine their appropriateness in
solving the recurrent food crisis in the region. Clearly the
development of the ASAL represents the highest poten-
tial for further economic advancement in the region. The
major challenge is how to utilize the available water––the
most limiting factor to economic activity in the dry areas.
Currently, most countries in the region are not able to
marshal financial resources to enable bulk water trans-
fers (e.g. inter-basin transfers), or dam and reservoir
construction for most of the ASAL. The pragmatic way
forward is in the development of least-cost small-scale
rainwater harvesting technologies by the communities
and individuals who live within these areas. Mere sur-
vival instinct has led many land users in the ASAL to
improvise various indigenous runoff-farming systems.
However, due to limited technical resources, these in-
digenous runoff-farming systems are poorly designed
and operated. Therefore, a great benefit can be realized
through technical improvements of the existing water
harvesting initiatives. This can be accomplished by first
understanding and evaluating the various systems being
used in the region and comparing their performance
vis-

a-vis the prevailing local conditions.
As water becomes more and more scarce, there is a
need for an integrated approach to water management
that encompasses all water users, types of water uses and
sources of water. Water management, however, can
never be an aim in itself, it is an integral part of farm
and land husbandry and its objective should always be
to protect and improve the land users’ situation (LEISA,
1998). Nevertheless, high-external-input techniques may
be too expensive for smallholders or are inappropriate
to local biophysical and social conditions. Many land
users would benefit from low-cost techniques more sui-
ted to their conditions and needs, and which also ensure
an increase in water use efficiency and conservation. It is
encouraging that land users have developed many low-
cost water saving techniques. Unfortunately, although
most of these innovations remain unrecognized, many of
them are within the reach of the land users. Therefore,
according to LEISA (1998), water scarcity can be chal-
lenged!
Traditional water harvesting systems are characterized
by flexibility and endurance and are strongly associated
with the people who live in marginal environments. Thus
different areas will have different techniques for harvest-
ing and applying water. Although the potential for water
harvesting has not been fully assessed, this potential is
probably quite large in the Greater Horn of Africa where
food security is a major concern. Recently, renewed in-
terest has been shown in water harvesting in sub-Saharan
Africa, probably as a result of increasing pressure on
land, which forces more and more people into dry areas
(Oweis et al., 1999). This new trend could also be at-
tributed to failure of more conventional methods and
changing environments forcing people to adopt new
survival strategies. Therefore, water harvesting has a
high potential for improving food security and reducing
over-dependency on food aid. However, for this poten-
tial to be realized, appropriate techniques need to be
identified for particular areas within the region. The case
studies will contribute towards identifying different
techniques land users in the region have already tested
and approved, and look into ways of improving the
adopted technologies.
Rainwater harvesting is a promising technology for
improving the livelihoods of many inhabitants of the
vast dry regions of the world. RWH can be viable in
areas with as low as 300 mm of annual rainfall (Kutch,
1982). However, Pacey and Cullis (1986) gave a more
conservative range of annual rainfall, 500–600 mm. But,
Kutch (1982) further stated that annual rainfall is not
the most important criterion. Nevertheless, the tech-
nology has been used to sustain food production in the
Negev desert of Israel with meagre annual rainfall of
about 100 mm (Shanan and Tadmor, 1976). Ironically,
most of the famine stricken areas of Africa receives
much more than 100 mm of rainfall.
Thus many parts of the SSA could tremendously
improve food security through RWH, which aim to
supply the deficit between rainfall and evapotranspira-
tion during the growing season. In case of RWH for
supplemental irrigation, the deficit is maintained by
supplying water to the crops during the critical periods.
Some experts regard irrigation as the only viable method
of agricultural production in the ASAL (Pacey and
Cullis, 1986). But history has proved otherwise espe-
cially for small scale farming systems. Therefore, pro-
motion of RWH should take into consideration the
perceived low rates of financial investments, especially in
 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
runoff farming, compared to irrigated agriculture. RWH
minimizes some of the problems associated with irriga-
tion such as competition for water between various uses
and users, low water use efficiency, and environmental
degradation. It is a simple, cheap and environmentally
friendly technology, which can be easily managed with
limited technical skills. The technology can also be in-
tegrated with many land use system, hence it is appro-
priate for local socio-cultural, economic and biophysical
conditions. Furthermore, there are many traditional
water management techniques still being used to make
optimal use of available rainfall (LEISA, 1998).
3.2. Reduction of conflicts over water resources
Extensive areas of the SSA countries are not well en-
dowed with water resources. This scarcity is aggravated
by poor distribution of water resources in most coun-
tries. For instance, in Kenya, less than 20% of the
country has adequate water resources for rainfed agri-
culture. In the vast dry areas, the main challenge is,
therefore, to increase water supply through more efficient
utilization of rainfall. It is evident that water scarcity is
one of the main drawbacks to substantial development of
the ASAL. This scarcity has led to persistent conflicts
over use and access to existing water supply. The con-
flicts involve different water users and uses.
More often than not, different clans especially within
the pastoral communities, in the ASAL have been en-
gaged in increasing conflicts over the control and use of
communal water sources and grazing land. Cross border
conflicts leading to severe clashes have also occurred
over control of natural resources. Notwithstanding ex-
isting traditional institutions, that to some extent have
promoted peaceful coexistence, the conflicts seem to get
worse by the day as water resources become scarcer.
Hence, one of the logical ways to contain the situation is
to provide adequate water and food supply. This ap-
proach has apparently worked well in northeastern
Kenya, especially in Wajir district, where a local NGO
has assisted in the construction of water pans to store
rainwater for different clans (Githinji, 1999). This is a
case where low technology––water harvesting––has
proved itself, not only as a water supply system, but also
as a conflict resolution mechanism. In addition, the
technology has led to improved food security and living
standards through provision of water for domestic,
livestock and agricultural purposes. This technology has
also created employment besides being easily replicable.
Similar cases will be articulated in the proposed project
and hence ways of dealing with the twin problem of food
security and conflicts over natural resources which is
prevalent in the GHA region.
Moreover, the case studies considered interrelated
environmental governance and gender issues affecting
food security and water availability. These include sus-
953
tainable environmental management strategies and tra-
ditional institutions that are involved in the well being of
the community and management of conflicts over use of
natural resources. Many of the conflicts, especially inter-
clan conflicts, are normally aggravated by food insecu-
rity and competition over scarce natural resources.
Conflicts over natural resources, especially water and
land, have been politicized in the SSA. According to
Mathenge (2002), the issue of water is of equal impor-
tance on the political scene as security in Laikipia Dis-
trict in upper Ewaso Ng’iro river basin. Large-scale
horticultural farming by wealthy local and international
concerns on the slopes of the Aberdares and Mt. Kenya
has depleted the mountain streams that used to be the
main sources of water leading to upstream–downstream
conflicts. Smallscale farmers along the streams have also
contributed to water conflicts by abstracting water, in
most case (more than 70%) illegally, for irrigation.
During extreme dry spells, the provincial administration
normally intervenes by banning water abstraction for
irrigation. Otherwise downstream users would organize
themselves and destroy water diversion structures up-
stream.
Insecurity too has contributed to the problem as
many farmers have abandoned livestock rearing––
attractive to cattle rustlers––to try out farming. This has
increased conflicts over water rights and food insecurity.
The politicians in the area have threatened to lead the
affected communities to storm horticultural farms over
water conflicts, while others have proposed that gov-
ernment impose levies on major horticultural producers
to raise funds to construct and maintain reservoirs to
harness flood waters. Hence RWH could play a major
role in conflict resolutions, especially in drier areas of
GHA. Some large-scale horticultural farmers have al-
ready adopted RWH by constructing large earthdams to
harvest runoff to supplement limited water for irriga-
tion. Another form of conflict occurs during the rainy
season over limited runoff on shared road/footpath
drainage. This is becoming common in Ng’arua division
of Laikipia District where neighbouring farmers com-
pete and some times fight over diversion of runoff to
their farms, especially those with farm ponds for storing
water for use during inter-seasonal droughts––to miti-
gate water stress during critical growth stages. This kind
of conflicts could be addressed through improved
management systems at community level.
4. Hydrological impacts: limits of up-scaling rainwater
harvesting
Rainwater harvesting involves abstraction of water
in the catchment upstream and may have hydrologi-
cal impacts on downstream water availability. Down-
stream access to water as a result of increased water
954 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
withdrawals upstream is an issue of concern, but it is
assumed that there are overall gains and synergies to be
made by maximizing the efficient use of rainwater at
farm level (Rockstr€ om, 1999). However, up-scaling of
RWH––increasing adoption––could have hydrological
impacts on river basin water resources management.
The on-going PhD study aims at assessing downstream–
upstream interaction related to increased adoption
rate––retaining more water in the watershed––in the
water scarce Ewaso Ngi’ro river basin in Kenya. Up-
grading rainfed agriculture, through the promotion of
RWH in the ASALs, require proper planning of land
management at river basin scale, rather than conven-
tional focus on farm level.
In the past, runoff has been as being destructive and
needed to be diverted from agricultural lands as wit-
nessed by over 30 years of soil conservation practices in
Kenya. However, radical transformation are required,
where surface runoff from upstream watershed entering
a farm will no longer be seen as a threat to be disposed
of or diverted away, but as a resource to be harnessed
and utilized to improve rainfed agriculture. Such
transformation is complex, especially among smallscale
farmers, since even a runoff from a small catchment will
involve multiple landusers. Presently there is little at-
tention given to ownership and management of locally
produced runoff, but this is expected to become a par-
amount issue if runoff is to be optimally managed on a
larger scale for local production purposes. In Laikipia
district of Kenya, conflict over runoff diversion and
utilization for crop production is a reality (Kihara,
2002). The situation may become much worse with the
growing realization of the benefits of RWH especially
for resource-poor smallscale farmers, who depend solely
on rainfed agriculture. Already even with limited RWH
systems, downstream–upstream conflicts between pas-
toralists and farmers (who divert meager stream water
for irrigation) are very common particularly during the
dry periods. The rainfed farmers are also in the course of
entering the conflict, among each other and with the
downstream landusers––both farmers and pastoralists in
this water-scarce river basin. The Indian experience on
communal rainwater management may provide useful
background in an attempt to develop sustainable RWH
up-scaling strategies in SSA.
In the view of ensuing competition and conflicts over
limited water resources, the hydrological impacts of up-
scaling RWH in a watershed/river basin need urgent
attention. The Indian experience, where the Rajastan
Irrigation Department, fearing that the communal RWH
structures would threaten water supply located down-
stream ordered its destruction (Agarwal et al., 2001),
indicates the need for policies, legislation and new insti-
tutional order to manage RWH, especially for agricul-
ture and livestock. In Kenya and many countries in SSA,
such policies and legislation are unfortunately lacking.
There is need for research to provide information to as-
sist decision and policy makers formulate sustainable
river basin water resources management strategies.
As shown by several hydrological studies at water-
shed and basin, upstream shifts in water flow parti-
tioning may result in complex and unexpected
downstream effects, both negative and positive, in terms
of water quality and quantity (Vertessy et al., 1996). In
general though, increasing the residence time of runoff
flow in a watershed, e.g., through RWH may have
positive environmental as well as hydrological implica-
tions/impacts downstream (Rockstr€ om et al., 2001). The
hydrological impacts at watershed/river basin level of
up-scaling system innovations, such as RWH, are still
unknown and require further research. The proposed
study aims to shed some light on this issue.
Increased withdrawals of water in rainfed and irri-
gated agriculture may have negative implications on
water availability to sustain hydro-ecological ecosystem
services. The expected shifts in water flows in the water
balance would affect both nature and economic sectors
depending on direct water withdrawals (Rockstr€ om
et al., 2001). Upgrading rainfed agriculture through
RWH that enables dry spells mitigation, would involve
the addition of water, through storage of runoff, to the
rainfed system. The cumulative effect of RWH may have
an impact on downstream water availability within a
river basin scale. The effects are bound to be site specific
and need to be studied further (Rockstr€ om et al., 2001).
The potential of developing small farm ponds and
earthdams for supplemental irrigation in SSA, is deter-
mined by a set of site specific biophysical and socio-
economic factors (Rockstr€ om, 2001), which include
practiced farming systems, population pressure, formal
and informal institutions, land tenure, economic envi-
ronment and social structures. Thus hydrological im-
pacts cannot be assessed in isolation. It is important to
analyze the downstream effects on water availability, for
example for household and livestock needs, as well as
heath and environmental impacts, before introducing a
technology which retains water upstream, an possibly
reducing river flows.
References
Agarwal, A., Narain, S., 1997. Dying Wisdom: The Rise, Fall and
Potential of India’s Traditional Water Harvesting Systems. Centre
for Science and Environment, Thomson Press Ltd, Faridabad.
Agarwal, A., Narain, S., Khurana, I., 2001. Making Water Every-
body’s Business: Practice and Policy of Water Harvesting. Centre
for Science and Environment (CSE), India.
Barron, J., Rockstr€om, J., Gichuki, F., 1999. Rain water management
for dry spell mitigation in semi-arid Kenya. E. Afr. Agr. Forestry J.
65 (1), 57–69.
Benites, J., Chuma, E., Fowler, R., Kienzle, J., Molapong, K., Manu,
J., Nyagumbo, I., Steiner, K., vanVeenhuizen, R. (Eds.), 1998.
 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956
Conservation tillage for sustainable agriculture. In: Proc. Interna-
tional Workshop, Harare, 22–27 June 1998. Part 1 (Workshop
Report). GTZ, Eschborn, Germany, p. 59.
Critchley, W., 1987. Some lessons from water harvesting in sub-
Saharan Africa. Report from a workshop held in Baringo, Kenya
on 13–17 October 1986. The World Bank eastern and southern
Africa Projects Department, p. 58.
Critchley, W., Siegert, K., 1991. Water harvesting: a manual for the
design and construction of water harvesting schemes for plant
production. FAO, AGL/MISC/17/91. Available from <http://
www.fao.org/docrep/U3160E00.htm>.
Cullis, A., Pacey, A., 1992. A development dialogue: rainwater
harvesting in Turkana. Intermediate Technology, London, UK,
p. 126.
Duckham, A.N., Masefield, G.B., 1985. Farming Systems of the
World. Praeger Publishers, New York, USA.
Evanari, M., Shanan, L., Tadmor, N.H., 1971. The Negev, The
Challenge of a Desert. Harvard University press, Cambridge, MA,
USA.
FAO, 1986. African Agriculture: The Next 25 Years. FAO, Rome,
Italy.
Fentaw, B., Alamerew E., Ali, S., 2002. Traditional rainwater
harvesting systems for food production: the case of Kobo Wereda,
northern Ethiopia. GHARP case study report. Greater Horn of
Africa Rainwater Partnership (GHARP), Kenya Rainwater Asso-
ciation, Nairobi, Kenya.
Fox, P., Rockstr€ om, J., 2000. Water harvesting for supplemental
irrigation of cereal crops to overcome intra-seasonal dry spells in
the Sahel. Phys. Chem. Earth (B) 25 (3), 289–296.
Githinji, J.M., 1999. Personal communication. He was the Project
Manager of Joint Relief and Rehabilitation Servives (JRRS), a
local NGO operating in northeastern Kenya.
Gould, J., Petersen, E.N., 1991. Rainwater Catchment Systems for
Domestic Supply: Design, Construction and Implementation. IT
Publications Ltd., London, UK.
Hai, M., 1998. Water harvesting: An illustrative manual for develop-
ment of microcatchment techniques for crop production in dry
areas. RELMA, Technical Handbook No. 16, Signal Press Ltd.,
Nairobi, Kenya.
Hatibu, N., Mahoo, H.F., Kajiru, G.J., 2000. The role of RWH in
agriculture and natural resources management: from mitigation
droughts to preventing floods. In: Hatibu, N., Mahoo, H.F. (Eds.),
Rainwater Harvesting for Natural Resources Management: A
Planning Guide for Tanzania. Technical Handbook No. 22.
RELMA, Nairobi, pp. 58–83.
Hunt-Davis, 1986. Agriculture, food, and the colonial period. In:
Hansen, A., McMillan, D.E. (Eds.), Food in Sub-Saharan Africa.
Rynne Rienner, Boulder, CO, USA, pp. 151–168.
Isika, M., Mutiso,G.C.M., Muyanga, M., 2002. Kitui sand dams and
food security. MVUA GHARP Newsletter, vol. 4, April 2002
Edition. Kenya Rainwater Association, Nairobi.
Jodha, N.S., Mascarenhas, A.C., 1985. Adjustment in Self-Provision-
ing Societies, SCOPE 27. John Wiley and Sons, New York, USA.
Kiggundu, N., 2002. Evaluation of rainwater harvesting systems in
Rakai and Mbarara Districts, Uganda. GHARP case study report.
Greater Horn of Africa Rainwater Partnership (GHARP), Kenya
Rainwater Association, Nairobi, Kenya.
Kihara, F.I., 2002. Evaluation of rainwater harvesting systems in
Laikipia District, Kenya. GHARP case study report. Greater Horn
of Africa Rainwater Partnership (GHARP), Kenya Rainwater
Association, Nairobi, Kenya.
Kijne, J., 2000. Water for food for sub-Saharan Africa. Article
prepared for FAO. Available from <http://www.fao.org/landand-
water/aglw/webpub/watfood.htm>.
Kutch, H., 1982. Principle features of a form of water concentrating
culture. Trier Geographical Studies No. 5. Trier, Germany.
955
Lameck, P., 2002. Evaluation of rainwater harvesting systems in
Dodoma District, Tanzania. GHARP case study report. Greater
Horn of Africa Rainwater Partnership (GHARP), Kenya Rain-
water Association, Nairobi, Kenya.
LEISA, 1998. Challenging water scarcity [Editorial]. ILEIA Newslett.
Low Ext. Input Sustain. Agri. 14 (1).
Lundgren, L., 1993. Twenty years of soil conservation in Eastern
Africa. Regional Soil Conservation Unit (RSCU) Report No. 9,
Majestic Printing Works, Nairobi, Kenya, p. 47.
Mathenge, G., 2002. Special report: Laikipia district constituency
profiles. Daily Nation, Monday, July 15, pp. 11–15.
Muni, R.K., 2002. Evaluation of rainwater harvesting systems in
Machakos District, Kenya. GHARP case study report. Greater
Horn of Africa Rainwater Partnership (GHARP), Kenya Rain-
water Association, Nairobi, Kenya.
Mwakalia, S.S., Hatibu, N., 1992. Rain water harvesting for crop
production in Tanzania. In: Proc. 3rd Annual Conf. SADC. Land
and Water Management Research Program, pp. 513–525.
Ngigi, S.N., 2001. Rainwater harvesting supplemental irrigation:
promising technology for enhancing food security in semi-arid
areas. In: Proceedings 10th IRCSA Conference, 10–14 September
2001, Mannheim, Germany.
Ngigi, S.N., Thome, J.N., Waweru, D.W., Blank, H.G., 2000. Low-
cost irrigation for poverty reduction: an evaluation of low-head
drip irrigation technologies in Kenya. IWMI Research Report.
Collaborative research project between University of Nairobi and
IWMI. IWMI Annual Report 2000–2001, pp. 23–29.
Ngure, K.N., 2002. Evaluation of rainwater harvesting systems in
Kitui District, Kenya. GHARP case study report. Greater Horn of
Africa Rainwater Partnership (GHARP), Kenya Rainwater Asso-
ciation, Nairobi, Kenya.
Oweis, T., Hachum, A., Kijne, J., 1999. Water harvesting and
supplemental irrigation for improved water use efficiency. SWIM
Paper No. 7. IWMI, Colombo, Sri Lanka, p. 41.
Pacey, A., Cullis, A., 1986. Rainwater Harvesting: The Collection of
Rainfall and Runoff in Rural Areas. IT Publications, SRP, Exeter,
London, UK.
Pwani, N.B., 2002. Personal communication. He is the Regional
WatSan Program officer, International Federation of Red Cross
and Red Crescent Societies, Regional Delegation Nairobi, Kenya.
Reij, C., Scoones, I., Toulin, C., 1996. Sustaining The Soil: Indigenous
Soil and Water Conservation in Africa. EarthScan, London, UK.
p. 228.
Rockstr€ om, J., 1999. On-farm green water estimates as a tool for
increased food production in water scarce regions. Phys. Chem.
Earth (B) 24 (4), 375–383.
Rockstr€ om, J., 2000. Water resources management in smallholder
farms in eastern and southern Africa: an overview. Phys. Chem.
Earth (B) 25 (3), 275–283.
Rockstr€ om, J., 2001. Green water security for the food makers of
tomorrow: windows of opportunity in drought-prone savannahs.
Water Sci. Technol. 43 (4), 71–78.
Rockstr€ om, J., Kitalyi, A., Mwalley, P., 1999. Conservation tillage and
integrated land management: field experiences and extension
approaches. Paper presented at the ATNESA/SANAT Interna-
tional Workshop on Empowering Farmers through Animal Trac-
tion into the 21st Century. South Africa, September 20–24, 1999.
Rockstr€ om, J., Barron, J., Fox, P., 2001. Water productivity in rainfed
agriculture: challenges and opportunities for smallholder farmers in
drought-prone tropical agro-systems. Paper presented at an IWMI
Workshop, Colombo, Sri Lanka, November 12–14, 2001.
Shanan, L., Tadmor, N.H., 1976. Micro-catchment systems for arid
zone development. Hebrew University of Jerusalem and Ministry
of Agriculture, Rehovot, Israel.
SIWI, 2001. Water harvesting for upgrading rainfed agriculture.
Problem analysis and research needs. SIWI Report No. 11.
956 S.N. Ngigi / Physics and Chemistry of the Earth 28 (2003) 943–956

Stockholm International Water Institute (SIWI), Stockholm,
Sweden, p. 97.
Thomas, D.B. (Ed.), 1997. Soil and Water Conservation Manual for
Kenya. SWCB, Ministry of Agriculture and Livestock Develop-
ment and Marketing, English Press, Nairobi, Kenya, p. 294.
Tiffen, M., Mortimore, M., Gichuki, F.N., 1994. More people less
erosion: Environmental recovery in Kenya. ACTS Press, Africa
Centre for Technology Studies, Nairobi, Kenya, p. 301.
UNDP/UNSO, 1997. Aridity zones and dryland populations: an
assessment of population levels in the world’s drylands with
particular reference to Africa. UNDP Office to Combat Deserti-
fication and Drought (UNSO), New York, USA.
Vertessy, R.A., Hatton, T.J., Benyon, R.G., Dawes, W.R., 1996.
Long-term growth and water balance predictions for a mountain
ash (Eucalyptus regnans) forest catchment subject to clearing-felling
and regeneration. Tree Physiol. 16 (1/2), 221–232.